previously I asked a comparable question: cx_Freeze unable fo uncover mkl: MKL deadly ERROR: Cannot fill mkl_intel_thread.dll

But now I have actually a ethereal difference. I want to operation the regime without installation anaconda, just within a cmd.exe terminal, but it seems I am doing something wrong or the it is no possible.

After generating my application through python setup.py bdist_msi utilizing cx-freeze, ns am able come install and then operation it in ~ an anaconda environment, yet if I simply open a cmd.exe terminal and run it, i get

INTEL MKL ERROR: The mentioned module could not it is in found. Mkl_intel_thread.dll.Intel MKL deadly ERROR: Cannot fill mkl_intel_thread.dll.However, once running

where mkl_intel_thread.dllthe dll is found, so ns think this method it"s registered in the system (I am much more used to usage Linux, so may be ns am wrong).

You are watching: Intel mkl fatal error: cannot load mkl_intel_thread.dll.

How could I fix this problem?


python dll cx-freeze intel-mkl
re-publishing
follow
edited Aug 20 "19 in ~ 7:34
*

jpeg
2,09233 yellow badges1616 silver badges2828 bronze badges
request Aug 20 "19 in ~ 6:35
*

ElBaulPElBaulP
5,30766 gold badges3333 silver badges7171 bronze title
1
include a comment |

4 answer 4


active oldest Votes
3
Maybe another DLL important for MKL, such together libiomp5md.dll for example, is absent and reasons the error. Watch Cannot fill mkl_intel_thread.dll on python executable, mine answer there and also its comments.

If this tho does not fix your problem, try to manually copy other DLLs indigenous the anaconda environment"s library path right into the application installation directory and also its lib subdirectory. Once you have found which exposed is missing, you have the right to use the include_files option of cx_Freeze come automatize this step in the setup (as friend know).

See more: B&Amp;B Brooklyn New York - B Impact Assessment: Measure What Matters Most

Another possible issue would certainly be that you have an incompatible variation of MKL installed on your system and also that the frozen applications finds this dorn one, yet this is unlikely unless you have actually a 32-bit Python environment on a 64-bit mechanism or have actually installed the applications on another system.

EDIT:

It might still additionally simply be that the frozen applications does not uncover mkl_intel_thread.dll although where finds it. Wherein looks in the system search path provided by the PATH atmosphere variable, while Python looks in the modules search path provided by sys.path, which typically does not encompass the content of PATH, watch Where is Python's sys.path initialized from? however on windows there is a fallback system for registered DLLs (I don"t know exactly how it works). Anyway, one should not depend on this fallback as soon as one intends come install and run the applications on one more system, since the crucial DLL might not be mounted there. For this reason the essential dependencies should constantly be had in the installation directory.